Construção e validação de listas de monossílabos para a realização do índice percentual de reconhecimento de fala

Abstract

The present research aimed to develop and validate recorded lists of monosyllabic words for obtaining the percentage of speech recognition rate and get evidence of content validity and construct for new lists. In the first stage of this research, 125 monosyllables of the Brazilian Portuguese language were selected, with different syllabic structure and analyzed by experts and not experts about their familiarity. After, the selected words were organized into five lists with 25 monosyllables in each list and recorded digitally, as determined by ISO 8253-3: 2012. Then the lists were presented with 40 dB sensation level in 40 normal-hearing subjects, right-handed, 18-44 years old, of different education levels, divided into four groups, so that they performed an auditory assessment, repeating the words heard. A word produced with more than one error has been replaced and one of the lists was deleted, leaving four lists made up of familiar and validated words about the content. In the second stage, the lists validated in the previous step were applied to 60 normal-hearing subjects, right-handed, 18-24 years old, to investigate the equivalence of the lists and the homogeneity of each list in relation to speech intelligibility obtained by the percentage of arrangement of words per list. The strategy of requested submission and response was the same as the previous step, but at this stage the chosen presentation level sought a variation of scores between 40% and 60%, which was obtained with the presentation of lists of words in the presence of ipsilateral speech noise type, set at 30 dB hearing level, with a signal/noise ratio of -1 dB. On the third stage, it the correlation between the speech recognition scores with new lists was examined, which are considered equivalent in the previous step, and the scores obtained by applying the Pen list and Mangabeira-Albernaz (1973), which is widely used in clinical routine evaluation. The lists were applied in 20 subjects with moderate and moderately severe hearing loss, considering 36 ears (18 right and 18 left). the recognition index of speech on the comfort level was obtained. All the words sent for assessment were considered familiar by examining the Content Validity Ratio when analyzing the results found. There was no statistically significant difference between the mean percentage of correct words per list. There was a statistically significant difference between the scores of the subjects in the list, resulting in two equivalent lists, an intermediate list and a completely different list. There was a strong positive linear correlation between the scores of the subjects with hearing loss obtained with the new lists of equivalent monosyllables and the lists of Pen and Mangabeira-Albernaz It was concluded that the content of four lists of monosyllables were developed and validated, and among these, two were considered equivalent and equally difficult among the items, and are strongly correlated to the lists produced by Pen and Mangabeira-Albernaz, offering evidence of content validity and construct for the proposed instrument.